Geek ~ Cinephile ~ #Bitcoin 🚀 ~ (6,9)
The Milky Way
Joined July 2016
- Tweets9.9K
- Following3.5K
- Followers311
- Likes47.2K
Hattori Hanzō retweeted
this is f*cking insane... Boris Cherny's team at Anthropic just shared 94 Claude plugins and skills for free. ALL OF THEM
one answers your emails. one chases clients who didn't pay. one finds the trap in your contract. one fixes your code while you sleep
55 plugins for Claude Code
20 plugins for real jobs: sales, legal, finance, marketing
19 skills: Excel, PowerPoint, PDF, Word and more
269,000 GitHub stars between them. agencies charge $300-1,500 to build just one of these. here all 94 cost $0
how to grab one:
1. open Claude Code and type /plugin
2. pick the one for the job you hate most
3. hit install
4. give it the first task in chat
plugins: github.com/anthropics/claude…
work plugins: github.com/anthropics/knowle…
skills: github.com/anthropics/skills
start small: one plugin today for the task you repeat every week
plugins are the start. how to build your own AI agents is in the article below ↓
Hattori Hanzō retweeted
why is it always just a screenshot?
can i see something functioning in the real world? in a product? something that is best in class?
passing EMC? manufacturable? sourcable with supply chains? in volume?
Hattori Hanzō retweeted
⏳ Last 10 hours to get the world's first 10Gbps USB 3.2 logic analyzer for $149, with AI MCP support 🤖 and mobile use on your phone 📱!
🏭 #SLogic32U3 is in full production now, and shipping starts this month! 🚀
kickstarter.com/projects/sip…
Hattori Hanzō retweeted
This is f*cking gold, the CEO of Obsidian put his own second brain on GitHub for free and people still build theirs from zero
7 free repos from the Obsidian team:
-> kepano obsidian (by kepano)
github.com/kepano/kepano-obs…
the vault template of the Obsidian CEO himself, copy it and skip the blank page
-> jsoncanvas (by obsidianmd)
github.com/obsidianmd/jsonca…
the open canvas format, your maps stay plain files that are yours forever
-> obsidian releases (by obsidianmd)
github.com/obsidianmd/obsidi…
every community plugin and theme in one list
-> obsidian sample plugin (by obsidianmd)
github.com/obsidianmd/obsidi…
the official template, build the plugin you are missing
-> 40 questions (by kepano)
github.com/kepano/40-questio…
40 questions that turn your notes into a yearly review
-> obsidian minimal (by kepano)
github.com/kepano/obsidian-m…
a clean theme so your vault stops looking like a mess
-> flexoki (by kepano)
github.com/kepano/flexoki
the color scheme he made for writing and code
your second brain should never start from a blank page
copy his vault and make it yours
Hattori Hanzō retweeted
This should not be something an AI already knows how to do.
An AI took a real Wi-Fi and Bluetooth dev board and rebuilt the entire thing from scratch while I watched, from the schematic to the last trace.
43 parts, 799 trace segments, 88 vias, two sides, on a board smaller than a matchbox.
It didn’t just dump parts onto the board. It put the chip on top with the antenna at the edge like it knew the signal needed room. Then it flipped the board over and packed 28 more parts on the back.
USB lines first, pins next, power last. That’s the exact order a senior engineer would follow, and nobody told it to.
Zero errors on its own rule check, and then it spun the finished board in 3D like it was mocking us.
People are still asking AI to draw cats, and meanwhile it’s learning to build the boards inside your gadgets.
Hattori Hanzō retweeted
It's kinda cool to see the models finally be able to do semi-serious stuff x-band phased-arrays
Hattori Hanzō retweeted
hey remember that time @i2cjak challenged me to an electronic design challenge and then he PROMPTLY pivoted to SW'e'?
i MERCIFULLY gave him 7 months to catch up… but this board is getting ordered this weekend.
Hattori Hanzō retweeted
DGX sparks near doubled since this, outperforming all datacenter gpus
4,800$ -> 9,500$ (and currently 8,200$)
closed source inference ipo vals crashed (anth 2Tn$ -> 1.3Tn$)
will only continue as more teams pivot to open models, early adopters gain both technical & asset moat
v4 flash 0731 and glm 5.3 flash are already match-to-beating opus 4.6 on a 2x sparks stack (less than 10k$) at home, ~90% inference speed of cloud anthro opus. quantized its even usable on a single spark (4k$), thats 1month api spend thats barely justifiable before for perf tradeoffs
v4.1 flash is now beating opus 5 and hostable on 3x sparks (less than 15k$) on day 1, with optimizations within a few weeks it wont even be close to where it is today which is already beating anth's flagship while being permanently hostable (zero api bills forever) for a single eng's payroll not to mention capacity for domain-specialized tuning
starting from today its basically a no brainer for the majority of funded orgs to not be on anth/oai anymore, the economics are just not justifiable
Hattori Hanzō retweeted
the NSA put its own hacking tool on GitHub and most people still haven't opened it
5 security repos that show you how exposed you actually are. all free, all open source.
want to see every account tied to your username across the internet?
-> Sherlock
github.com/sherlock-proje…
want to take any program apart and read what it really does?
-> Ghidra, the reverse engineering framework built by the NSA
github.com/NationalSecuri…
want to watch everything your phone apps send home?
-> mitmproxy
github.com/mitmproxy/mitm…
want to know if your API keys and passwords are already sitting in your git history?
-> TruffleHog
github.com/trufflesecurit…
want to test your own WiFi and Bluetooth with a $10 chip?
-> ESP32 Marauder
github.com/justcallmekoko…
~260k stars of tools that professionals get paid to run.
none of this is hidden. it's the same kit pentesters use on the job.
the catch: it's only legal on things you own or have written permission to test.
the best way to learn how you get hacked is to run the attack on yourself first.
bookmark this. pick one. point it at your own stuff tonight.
Hattori Hanzō retweeted
I am begging you to use these 2 GitHub repos - they're f*cking BROKEN...
one repo turns AI agents into a SELF-IMPROVING trading desk, the other gives them eyes to see the ENTIRE internet
i recorded 2 hours of the REAL BTC order book & built this loop from scratch
my setup scores 8 agents, rewrites the worst one, kills the losers & spawns new ones on its own (69 REWRITES, 30 SURVIVED)
here is the EXACT build from scratch:
1. open Claude Code in an empty folder & paste: "read architecture/autoresearch.md in github.com/chrisworsey55/atl… and explain the loop in five lines"
2. paste: "install github.com/Panniantong/Agent… in a virtualenv and list the channels that work without a login"
3. paste: "record the Kraken BTC/USD book, 300 levels per side, and every trade from the public API for 2 hours"
4. paste: "build 8 agents in a git repo, each one a config file: lookback, block length, drift, vol scale, book imbalance on or off"
5. paste: "each agent resamples real 1-minute moves into 400 futures, using only data before each decision"
6. paste: "score them walk-forward on pinball loss at +15 minutes"
7. paste: "every round the worst agent gets one change. if its score improves, git commit. if not, git reset --hard"
8. paste: "weights start at 1.0. top quartile x1.05, bottom quartile x0.95, floor 0.3, cap 2.5. an agent at the floor for 8 rounds goes extinct, and a new one spawns when the mix shows a gap"
9. paste: "mix the survivors by weight into 1,000 futures from NOW, print the share above NOW, and stack them as a 3D ridge next to the order book"
i won't say anything more:
Hattori Hanzō retweeted
BlackRock just leaked free repo that should be BANNED...
Jane Street ($39.6B) and Citadel ($12.2B) don't guess price, they take the spread. Now BlackRock's optimizer + spread bot + Drex do it for Claude Code
here's full how to setup it:
1. Hummingbot (21K stars) - your own mini Jane Street: it places a buy just below the price and a sell just above it, and earns the gap 24/7
2. HOLA (by BlackRock) - BlackRock's own open-source optimizer on a Rust engine: it tries hundreds of settings on past data and picks the spread, order size and limits for you
3. Drex (just launched) - a decision model built for finance: one question every minute, "keep quoting or pull out?", answered with a probability, not an essay
here's how to set it up:
1 → paste the 3 links into Claude Code (below) and say: "install all 3, Hummingbot in paper trade mode"
2 → grab your Drex key from @NaceAI at console.nace.ai (BODILA-1) and say: "check my Drex key"
3 → pick the market: "find Binance pairs where the spread is wider than 2x my maker fee, take the 3 with the most volume and download their last 30 days of trades"
most bots die right here: on BTC the spread is usually 1 cent and fees eat everything
4 → decide: "write a Hummingbot market-making script: every minute ask Drex if the market is CHOPPY, TRENDING or PANIC, quote only in CHOPPY, pull the orders otherwise. when inventory piles up, shift both quotes so it unloads first"
every minute becomes one question: stay or step aside? the answer is a probability, not an essay
5 → tune: "ask Drex once for every minute of history and cache it. let HOLA tune the spread, order size, inventory limit and the Drex threshold on the first 3 weeks. count a fill only when price trades through my quote. test the winner on the last week it never saw. if it loses after fees, change the pair"
6 → paper trade 2 weeks on live data: "run the winner on binance_paper_trade and log every fill and every Drex call". go live only if the test and paper both win after fees, and start with money you can afford to lose
the result: your own mini Jane Street that quotes 24/7, gets its numbers from BlackRock's optimizer and steps aside the moment Drex smells a trend
that's it.
Hattori Hanzō retweeted
Hardware engineers should probably look away now.
One command, and an AI agent built an entire circuit board from scratch while I watched. It drew the schematic, placed every component, routed every trace and checked its own work for errors.
44 components, 55 connections, 191 vias and 1442 mm of copper, done in under a minute.
The design rule check came back with zero errors and zero warnings, and then the agent powered the board on in 3D as if to say “your turn.”
This is a real board layout that’s ready to be manufactured.
Hattori Hanzō retweeted
THIS IS ACTUAL F*CKING GOLD
Anthropic just shared one of the best tricks for Claude Code.
The idea: run Sonnet, Haiku and Opus 5.5 together, each one doing only the job it's best at.
Opus-level judgment on every big call, without paying Opus rates for every routine turn.
You just launch Claude Code like this:
claude --model sonnet --advisor opus
here's the split:
→ Sonnet 5.5 runs the session and writes the code
→ Haiku 5.5 explores files and docs in parallel
→ Opus 5.5 stays on call as advisor for the big calls
→ you stop spending Opus on work that doesn't need it
Opus only steps in at three moments:
→ before a plan locks
→ when the same error shows up twice
→ before calling it done
Sonnet builds. Haiku explores. Opus reviews.
JEV engineering is the same move one layer down: the forks that need no thinker (which file, which tool, retry or stop) go to JEV in milliseconds, and the big models only see the ones that split
the full tree
> Sonnet 5.5 runs the main session and edits code
> search maps the repo on Haiku 5.5
> grep finds call sites on Haiku 5.5
> docs pulls the API reference on Haiku 5.5
> Opus 5.5 on call as advisor, reads the whole transcript when consulted
> JEV routes micro-forks in milliseconds
paste the tree and this prompt into Claude Code ↓
"Rebuild my Claude Code setup around this tree:
1. Run the main session on Sonnet 5.5 with Opus 5.5 as advisor.
2. Create three read-only Haiku 5.5 subagents: search (maps the repo), grep (finds call sites), docs (pulls the API reference).
3. Add one rule to CLAUDE.md: consult the advisor before locking a plan, after the same error twice, and before calling a task done.
4. Route fast deterministic forks (tool choice, path selection, retries) through JEV.
Show me every change as a diff first. No edits until I say go."
↳ code.claude.com/docs/en/advi…
save it, you'll use this every single day.
Hattori Hanzō retweeted
ESTO ES RELAMENTE P*TAMENTE PELIGROSO...
el creador de REA acaba de soltar 20 repos de GitHub para montarte un p*to estudio de ingeniería inversa:
[01] núcleo de análisis
> REA: un único MCP que se conecta a todo github.com/morluto/rea
> Ghidra: decompilador, gratis github.com/NationalSecuri…
> GhidraMCP: deja que tu agente maneje Ghidra github.com/LaurieWired/Gh…
> ida-pro-mcp: conecta tu agente a IDA Pro github.com/mrexodia/ida-p…
[02] móvil y apps
> jadx: convierte APKs de Android de vuelta a Java github.com/skylot/jadx
> Apktool: desmonta APKs y vuelve a montarlas github.com/iBotPeaches/Ap…
> Frida: engánchate a las apps mientras se ejecutan github.com/frida/frida
> ILSpy: decompila .NET github.com/icsharpcode/IL…
[03] binarios y bajo nivel
> radare2: framework de RE que vive en tu terminal github.com/radareorg/rada…
> x64dbg: depurador para Windows github.com/x64dbg/x64dbg
> ImHex: editor hexadecimal con su propio lenguaje de patrones github.com/WerWolv/ImHex
> angr: motor de ejecución simbólica github.com/angr/angr
[04] red y ejecución
> mitmproxy: se pone en medio y lee el tráfico github.com/mitmproxy/mitm…
> dnSpy: depura y edita apps .NET github.com/dnSpyEx/dnSpy
> rizin: el fork de radare2 github.com/rizinorg/rizin
> raddebugger: el depurador de Epic Games github.com/EpicGames/radd…
[05] reconstruir y verificar
> awesome-ai-reverse-engineering: 122 repos, todos puntuados github.com/zhuyansen/awes…
> awesome-reverse-engineering: seleccionado por el creador de REA github.com/morluto/awesom…
> flujos de REA: prompts de investigación listos para usar, vienen con REA
> bucle clean-room: SPEC.md, oráculos y comprobaciones hasta que coincida
cómo funciona: binario -> REA -> SPEC.md -> reconstruir -> oráculo -> reintentar
guárdate esto, apúntalo a cualquier app y mira cómo se explica sola.
Hattori Hanzō retweeted
this is getting really dangerous...
REA creator posted 20 GitHub repos to build a whole reverse engineering studio:
[01] analysis core
> REA: a single MCP that plugs into everything github.com/morluto/rea
> Ghidra: decompiler, free github.com/NationalSecuri…
> GhidraMCP: lets your agent drive Ghidra github.com/LaurieWired/Gh…
> ida-pro-mcp: hooks your agent into IDA Pro github.com/mrexodia/ida-p…
[02] mobile & apps
> jadx: turns Android APKs back into Java github.com/skylot/jadx
> Apktool: take APKs apart and put them back together github.com/iBotPeaches/Ap…
> Frida: hook into apps while they run github.com/frida/frida
> ILSpy: decompiles .NET github.com/icsharpcode/IL…
[03] binary & low-level
> radare2: RE framework that lives in your terminal github.com/radareorg/rada…
> x64dbg: debugger for Windows github.com/x64dbg/x64dbg
> ImHex: hex editor with its own pattern language github.com/WerWolv/ImHex
> angr: symbolic execution engine github.com/angr/angr
[04] network & runtime
> mitmproxy: sits in the middle and reads the traffic github.com/mitmproxy/mitm…
> dnSpy: debug and edit .NET apps github.com/dnSpyEx/dnSpy
> rizin: the radare2 fork github.com/rizinorg/rizin
> raddebugger: the debugger from Epic Games github.com/EpicGames/radd…
[05] rebuild & verify
> awesome-ai-reverse-engineering: 122 repos, every one graded github.com/zhuyansen/awes…
> awesome-reverse-engineering: curated by the guy who made REA github.com/morluto/awesom…
> REA workflows: ready-made investigation prompts, ship with REA
> clean-room loop: SPEC.md, oracles and checks until it matches
how it runs: binary -> REA -> SPEC.md -> rebuild -> oracle -> retry
save this, point it at any app and watch it explain itself, then read the full guide below:
Hattori Hanzō retweeted
eSIM卡管理项目 vohive作者又出手了
vohive 2 ，这图真带劲
vohive是eSIM卡免手机收发短信及电话的开源项目
因一些原因原作者删库了
最近放出了vohive 2 的demo⬇️
很悬！功能强大了更多！
期待完整项目放出！
如果手里一堆eSIM，如何免手机收发短信&保号办法⬇️
在 Mac（Apple Silicon / Intel）上用 UTM 跑 Linux 虚拟机
把大疆 4G 模块（EG25-G）伪装成移远 Quectel EC25
并部署 vohive 平台的完整步骤
github.com/wlzh/dji-4g-vohiv…
Hattori Hanzō retweeted
Alright 0-1 on the start of NHL, but determined to have another streak like the World Cup.
Taking Boston over Pitt tonight.
🤝 Paid partnership
Hattori Hanzō retweeted
Run your firmware before you order the PCB.
We built firmware simulation into HeyPCB. Click buttons on your board’s 3D model and watch the screen respond, right in your browser.
heypcb.ai/dashboard
Hattori Hanzō retweeted
I’m doing this with Codex + KiCad for MOSS, my litter-picking robot. It’s my first PCB: power distribution, servo control, audio and current sensing for both track motors, all on one board.
I’ve submitted it to @PCBWayOfficial for review. Posting the design as I went brought in some really useful advice, and I realised I could simplify the whole wiring harness too.
The boards haven’t arrived yet, so I can’t claim it works. But being able to get this far is pretty exciting for someone whose first job was graphic design!
